[ad_1]
On Tuesday Professor Neil Ferguson stepped down as an adviser to the British government. His married lover had been with him twice, despite the fact that she lives in another part of London, and despite the fact that his own calculations were behind the rules prohibiting such visits since mid-March.
The UK government changed the strategy and imposed severe restrictions on the public due to a report by Neil Ferguson’s research team at Imperial College. According to him, around 250,000 people would die and attention would be overburdened if the country did not enter to completely stop the covid-19 epidemic, instead of limiting it. One of the measures was to reduce physical social contacts between people as much as possible, including by prohibiting couples who do not live together from meeting. Ferguson leaves because he has now clearly violated his own recommendations. There have also been serious criticisms of his model, which, according to anonymous programmers, contains significant problems and errors.
The Imperial College report It received a great deal of attention around the world, and several Swedish research groups used models like his to make forecasts of developments in Sweden. On April 7, Joacim Rocklöv, professor of epidemiology at Umeå University, and his staff published their results on MedRxiv, an alleged preprintserver, a database where researchers can see the work of others before being reviewed by other experts, approved and published.
In their model, the researchers have based on the age distribution of the population in all municipalities in Sweden, at what proportion of the different age groups that can be infected by the virus, are exposed to the infection, become infected and they are recovering. They have also calculated how people travel, what care needs patients have and how much it will cost.
– In the end, it gets pretty complicated. There are so many parameters The parameter values have been derived from the scientific literature, for example from the Ferguson model, but they have also made their own estimates. As I understand it, they have not adjusted the parameters to match the model as best as possible with real data, but instead verified that the model provides reasonable forecasts for historical data, says Philip Gerlee, associate professor of biomathematics at Chalmers University of Technology and Gothenburg University.
The conclusions of the report are serious for the country. Only if the truly strenuous efforts to completely stop the spread of the infection had been introduced since March 20, medical care could meet the need for intensive care sites, according to their model. However, without additional restrictions to limit the spread of the infection, 16,000 people would need intensive care now in early May. In an intermediate scenario with slightly stricter measures, which had limited social contacts for anyone under the age of 60 with a quarter and for all those over 60 to half, the need for intensive care places would still be between 5,000 and 9,000 when it is highest in June.
In reality, the number of covid-19 patients receiving intensive care has been fairly constant between 500 and 550 since mid-April.
– So, even when they believe in a peak of between 5,000 and 9,000 intensive care units in May and June, they end up with a factor of ten errors, says Philip Gerlee.
April 15 Another calculation for Sweden was published on MedRxiv by, among others, Lynn Kamerlin, professor of structural biology at Uppsala University. The model that the researchers use is even more complex: they have calculated each individual in Sweden and how they move between homes, work and other places in society.
“It becomes extremely complicated and requires many parameters to describe the spread of the infection,” says Philip Gerlee.
Philip Gerlee, who works with mathematical models for disease, is one of the authors of a debate article in the Ny Teknik newspaper, with strong criticism of Lynn Kamerlin’s model colleagues. The criticism is especially true that the model contains more than 100 parameters and that its value does not come from actual measurements.
– I reacted saying that they are contagious even in hospitals, but reduced to 25 percent. Therefore, when healthcare personnel encounter a seriously ill person entering intensive care, they expect that there is a fairly high risk of infection. And they conclude that with today’s measures, 90 percent of all hospital staff would be infected in early May. That is not reasonable, says Philip Gerlee.
That is not the only unreasonable conclusion in the report. The researchers found that with Sweden’s current strategy, 96,000 people would have died in covid-19 before July 1. This corresponds to more than 1,000 deaths per day during the spring from the time the report was released in April, to compare with the total number of deaths, which is now just over 3,000 people.
With the current strategy, more than 20,000 covid-19 patients would also require intensive care in early May, according to estimates. If Sweden had imposed the strictest restrictions to limit the spread of the infection since April 1, 7,000 Swedes would still need intensive care in late April, and more than 20,000 die before July 1, according to Lynn Kamerlin and her peer model. of work.
– My impression is that the model was developed more as propaganda than to describe the spread of infection. Even when the article was published, the deviations were too large compared to the available data, says Philip Gerlee.
Imperial College also clearly distances itself from Twitter modeling on Twitter.
https://twitter.com/imperialcollege/status/1257991340505346048
The Public Health Authority has not made any model for development in the country, but has prepared a planning base for the care to be prepared. According to its worst case scenario, at most around 1,700 people may need intensive care at the same time. This is well above the current 500 patients, but well below the model’s predictions.
Joacim Rocklöv He tells DN that a new version of the report will be published soon on MedRxiv.
– Since the new data comes from the Swedish situation, we have been able to adjust the model to better describe the situation. Furthermore, during this time that has passed, it is clear that social distancing has had a great impact. Among other things, we wanted to show that there was an opportunity to influence and avoid the high-risk scenario with the help of social distance, “he says.
Lynn Kamerlin believes that her model still stands.
– I think, above all, we should be glad that the changing behavior of people made us avoid the scene of disaster. It is really sad how people talk about this. They see it the wrong way. Instead of asking ourselves why this did not happen, we should celebrate. Because it could have been, she says.
[ad_2]
